FIR #462: Cheaters Never Prosper (Unless They’re Paid $5 Million for Their Tool)

FIR #462: Cheaters Never Prosper (Unless They're Paid $5 Million for Their Tool)

A Columbia University student was expelled for developing an AI-driven tool to help applicants to software coding jobs cheat on the tests employers require them to take. You can call such a tool deplorable or agree with the student that it’s a legit resource. It’s hard to argue with the $5 million in seed funding the student and his partner have raised. Also in this long-form monthly episode for April 2025:

How communicators can use each of the seven categories of AI agents that are on their way.
LinkedIn and Bluesky have updated their verification programs in ways that will matter to communicators.
Onboarding new talent is an everyday business activity that is in serious need of improvement.
A new report finds significant gaps between generations in the PR industry when it comes to the major factors impacting communication.
Anthropic—the company behind the Claude LLs—warns that fully AI employees are only a year away.
In his Tech Report, Dan York explains how Bluesky experienced an outage even though they’re supposed to operate under a distributed model.

Links from this episode

Links from Dan York’s Tech Report

The next monthly, long-form episode of FIR will drop on Monday, May 26.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email .(JavaScript must be enabled to view this email address).

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. Shel has started a metaverse-focused Flipboard magazine. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville Hobson: Greetings everyone, and welcome to for immediate release episode 462, our monthly long form edition for April, 2025. Neville Hobson in.

Shel Holtz: I’m Shell Holtz in Concord, California in the us. We’re thrilled to be back to tackle six topics that we think communicators and others in business will find interesting and useful.

Before we jump into those topics, though, as usual, in our monthly episode, we’d like to recap the shorter episodes that we’ve recorded since the last monthly, and we’re. Neville over. I think we’re,

Neville Hobson (2): yeah, I think we are. Shell, uh, episode 4 56. That was our March monthly recorded on the 24th of, or rather, published on the 24th of March.

Um, a lot of topics in that one, they addressed variety of issues. Uh, for instance, uh, publishing platform ghost enabling the social web by employees quitting [00:01:00] over poor communication in companies, the UK newspaper launching AI curated news. And there were three or four other topics in there too. Plus Dan York’s tech report as usual.

So that’s a mighty episode. And.

Shel Holtz: We did on the topic of whether artificial intelligence will put the expertise of practice by communicators at risk. Julie MayT wrote, it’s not about what we do anymore, but how we think, connect and interpret. Human value isn’t disappearing. It’s shifting, isn’t it? The real opportunity isnt doubling down on creativity, context and emotional intelligence by communicating with kindness and empathy.

Looking forward to tuning in. And Paul Harper responded to that comment saying, my concern is that AI, for many applications completely misses emotional intelligence, cold words, which are taken from the web, which does not discriminate between good and bad sources, truth or fake. And Julie responded to that saying, good point, Paul.

When it comes to important [00:02:00] stuff where it really matters whether AI is giving us something real or fake, I usually ask for the source and double check it myself. Chachi PT also has a deep research function that can help dig a bit further.

Neville Hobson (2): Okay, so our next 1, 4 57 that was published on the 28th of March.

And this I found a, a really interesting discussion, very timely one, talking about communicating the impacts of Mr. Trump’s tariffs. And we talked about that at some length. Our concluding statement in that episode was communicated should counsel leaders on how to address the impacts of those tariffs.

And I believe we have a comment on that show

Shel Holtz: from Rick Murray, uh, saying So true business models for creative industries are being turned upside down, revenue and margin streams that once fueled agencies of all types don’t need to exist now and won’t exist in three years.

Neville Hobson (2): Well said Rick. Well said 58, which we recorded or published on the 3rd of April.

This was, I thought, a [00:03:00] really interesting one, and we’re gonna reference it again in this episode. This was about preparing managers to manage human AI hybrid. Teams, um, a lot of talk about that and that how, uh, uh, uh, that we are ready or not for this, it’s on the horizon. It’s coming where we will have this in workplaces, and we talked about that at some length in that episode.

Uh, looking at what it means for managers and how far businesses from, uh, how far it is from enabling their managers to succeed in the new work reality. We also added a, a kind of a, a mirror or a parallel element to this, that it’s also helping employees understand what this means to them in the workplace if they got AI colleagues.

So, um, I don’t think we had any comments to that one. She, but it’s got a lot of views, so people thought about that, just didn’t, didn’t have any comments at this point, but great topic. Uh, I think

Shel Holtz: left, left them speechless if we did.

Neville Hobson (2): Yeah, exactly. So, uh, maybe we’ll get some after this episode in nine that we publish on the 9th of April that [00:04:00] looked at how AI is transforming content from passive to interactive.

We discussed the evolving landscape of podcast consumption, particularly in light of Satya Nadal, the CEO of Microsoft, his innovative approach to engaging with audio content through ai. So not listening to the podcast, he has his, uh, chat bot of, uh, his favorite chat bot, not chat, GBT of course, it’s co-pilot that, uh, talks to the transcript and ge he engages that way.

Interesting. Uh, I’ve seen comments elsewhere about this, that, that say, why on earth do you wanna do this? But you can listen. Well, everyone’s got different desires and wishes in this kind of thing. Uh, but it seems to me a feasible thing to do it the, for the reasons he describes why he’s doing it. And I believe it attracted a number of comments.

Did it not show.

Shel Holtz: We did, starting with Jeff Deonna, who wrote, to be honest, I find this approach deeply disrespectful to podcast hosts and their guests. It literally silences their human voices in favor of a fake conversation with a solace [00:05:00] algorithm. Now, I responded to that. I thought that Cliff notes would be a reasonable analogy.

People rather than reading Silas Marner, uh, read the Cliff notes where some solace Summarizers outlines the story and tells you who the key characters are so that you can pass a test and it silences the voice of the author, author. And yet we didn’t hear that kind of objection to Cliff Notes. We’ve heard other objections.

Of course, you should read the whole damn book. Right? But I think people have been summarizing for years. Executives give reports to their admins and say, write me a one page summary of this. And now we’re just using. AI to do the same thing. I don’t know if you had any additional thoughts on Jeff’s comment.

Sure.

Neville Hobson (2): I left a comment to his, uh, comment. I just reply to his comment as well, saying that, uh, I didn’t say these words, but effectively it was a polite way of saying I disagree. Sorry, you’re not right with this for the reasons you’ve, you’ve outlined. I don’t have the comment open on my [00:06:00] screen now, so I can’t remember the exact words I used, but I thought I couldn’t let him get away with, with that, without a response.

Shel Holtz: Well, we had another comment from Kevin Anselmo, who used to do the Higher Education podcast on the FIR Podcast Network. He said, I asked chat GPT to summarize your podcast transcript. After receiving the below chat, GPT provided practical advice on actioning the takeaways in my own projects. Interesting exercise, and I will not read everything he pasted in from chat GT’s analysis of the transcript of our podcast.

But I’ll, I’ll tell you what the five key takeaway labels are. Transcripts are becoming essential. A ai AI makes podcasts interactive. Most people still prefer passive listening. AI is going multimodal. And then there’s a notable quote from the podcast, so that was, uh, turnabout. I mean, we’re talking about what would happen if people didn’t listen to the authentic voices.

Well, you know, Kevin didn’t have to listen to us. I’m fine with that. If he [00:07:00] walks away with actionable items based on hearing or reading a summary of our transcript, one more way to get to it. I agree. And Mark Hillary wrote, why would you need a transcript for chat GPT though? Just feed it the audio and it could work out what is being said.

Anyway, I.

Neville Hobson (2): Yeah, I replied to him as well. We had quite an interchange. I can’t remember if it was on LinkedIn or on on Blue Sky, I can’t remember which, which service now. Um, but um, he was gonna go and experiment himself with something else. Uh, ‘cause what he described, and someone else was left to comment about this as well.

Actually, I think that was on Blue Sky too, that, um, talked about, uh, you know, why would you wanna do this a bit bit like GE actually, not like Jeff. It wasn’t just alleging disrespect, it was saying, why would you wanna do this? Um, when I, you know, it was actually Mark who said he’d uploaded an MP three. And, uh, it had done the job.

It actually hadn’t, uh, chat. GPT got the MP three, created the transcript from it, and then it did what it [00:08:00] needed to do. So the transcript is essential to.

Shel Holtz: Whether you created Issa. Nevertheless,

Neville Hobson (2): these, these, yeah, these, these great comments are, are fab to have these I must have been extends the conversation.

Okay. So then four 60, which we published on April the 14th. This one talked about layoffs in the United States primarily, and the return of toxic workplaces and the big boss unquote era. Uh, the tide is turning. We started off and assessed that I mentioned. We’re seeing not, not the same and not layoffs per se, but people quitting here in the UK for different reasons.

But this turmoil in this and toxicity in the workplace is part of the reasoning. So we explore the reasons behind the layoffs in the US are the impact of CEO Tough talk and how communicators can help maintain a strong non-toxic workplace. So that was good. We have comments too, don’t we?

Shel Holtz: We do.[00:09:00]

Starting with Natasha Gonzalez who says something that stood out for me was a point that Neville made about employees in the UK who are resigning from jobs due to toxic workplace culture, rather than being laid off as in the us. I imagine this isn’t unique to the uk. And then Julie MayT, who was the first comment she’s going to bookend our comments, wrote that organizations in the US are starting to see we cracks in psychological safety and trust disappearing.

Then all those folks who keep everything ticking along will start to quietly disengage. It’s up to us, calms people to be brave enough and skilled to say on a wee minute, that message isn’t landing the way you think it is. While the big wigs are busy shouting, spinning, and flexing, it’s us who need to rock up with the calm, clear human communications, no drama, ram, just stuff that makes sense and actually help folks to figure out what the hell is [00:10:00] going on and what to do next.

Neville Hobson (2): Good comment Mr. Bit. And that takes us to the last one before this episode, episode 4 61. We published on the, on the 24th of April that looked at trends in YouTube video two reports in particular that really had interesting insights on virtual influences and AI generated videos. And the bit that caught my attention mostly was, uh, news that every video uploaded to YouTube.

So you take your video, you upload it, um, uh, can be dubbed into every spoken language on the planet, uh, with the, with the speaker’s lips reanimated to sync with the words they are speaking. I mean, this is either terrifically exciting or utter nightmare that, uh, that is approaching fast. So, um, we talked about that and uh, we haven’t had any comments to that one yet, but this is a topic I see I’m seeing quite a bit being discussed online in various places.

So this is just a start of this, I think. [00:11:00] So that takes us to the end of the recap show,

Shel Holtz: so I didn’t see it. Okay. Lemme talk about that.

Neville Hobson (2): And last but certainly not least, I want to mention a new interview that, uh, that we posted on the 23rd of April. This was with Zoa artists in Australia who we interviewed on an article she wrote in the populous blog on bridging AI and human connection in internal communication. It was a really, really good discussion we had with, uh, it’s definitely worth your time listening to this one.

You will learn quite a lot from what or Zoa has to say on this topic. What did you think of it? She, it was good, wasn’t it?

Shel Holtz: It was fascinating and I read that, that post in the popular blog and also was engaged in a conversation with Zuora at the Team Flow Institute where we’re both research fellows and she raised it and it led to a conversation with all the fellows [00:12:00] and this notion of what would a board of directors do if AI was in the room with them right now?

What would they use it for? How would they take advantage of it to some fascinating discussion. So worth a listen. Also up now is episode number 115 of Circle of Fellows, the monthly livestream panel discussion that people who watch live are able to participate in in real time. This was about communicating amidst the rise of misinformation and disinformation.

Brad Whitworth moderated this installment of Circle of Fellows with panelists, Alice Brink, Julie Holloway, and George McGrath. Sue Human was supposed to participate, but woke up feeling ill, but did send in some written contributions that, uh, were read into the discussion. So a good one. I’ve, I’ve listened to it.

You should too. It’s a very timely topic. And just to let you know about the next Circle, circle of Fellows, episode one [00:13:00] 16 is scheduled for noon eastern time on Thursday, May 22nd. The topic is moving to teaching. This is something a lot of communicators do is become adjunct professors or full professors, or even tenured professors.

And we’ll be having a conversation with four IABC fellows who have done just that, Cindy smi, John Clemens, mark Schumann, and Jennifer W. And in fact, I’m speaking at Jennifer W’s class via Zoom pretty soon, so that’ll be a fun one too. You can mark that one on your calendars May 22nd noon eastern time, and that’ll take us to the start of the coverage of our topics for this month, but only after we turn things over to an advertiser for a moment.[00:14:00]

As we have been discussing for some time, AI agents are coming and to a degree they’re already here. Ethan Molik, the Horton professor, and ai, I guess you’d call him an AI influencer. He posted this observation to LinkedIn a few days ago. He wrote, I don’t think people realize how much, even a mildly agentic AI system like chat PT oh three can do on its own.

For example, this prompt works in oh three zero shot. Come up with 20 clever ideas from marketing slogans for a new mail order. Cheese shop. Develop criteria and select the best one. Then build a financial and marketing plan for the shop, revising as needed, and analyzing competition. Then generate an appropriate logo using the image generator and build a website for the shop as a mockup.

Making sure to carry five to 10 cheeses to fit the marketing plan. With that single prompt in less than two [00:15:00] minutes, the AI not only provided a list of slogans, but ranked and selected an option, did web research, developed a logo, built marketing and financial plans, and launched a demo website for me to react to the fact that my instructions were vague and that common sense was required to make decisions about how to address them was not a barrier.

And that’s an open AI reasoning model, not an actual agent. Built to be an agent to take on autonomous tasks in sequence multiple tasks in pursuit of a goal with agents imminent. HubSpot shared a list of seven types of agents in a post on its blog, and I thought it would be instructive given what Professor Mooch wrote to, to go over these seven categories or classes of agents and where they intersect with what we do as communicators.

Now I, I’ll give you the caveat that. Somebody else may develop a different list. Somebody else may slice and dice the [00:16:00] types of agents differently, but this is the first time I’ve seen this categorization, so I thought it was worth going through. They start with simple reflex agents that operate based on direct condition action rules without any memory of anything that you may have interacted with it about before.

So in PR, we could use this for automated media monitoring alerts set up agents that trigger. Instant alerts based on keywords that, uh, appear in news articles or on social media that lets you respond quickly. Uh, you could have some basic chat bot responses, you right, simple chat bots on internal or external platforms that will answer frequently asked questions with pre-programmed answers about things like, I don’t know, office hours, basic company information, dates of upcoming events.

And then you could filter inbound communication, automatically flag or filter incoming emails or messages based on keywords that indicate urgency or specific topics and route [00:17:00] them to the appropriate team member to respond to it. The second type of agent is a model-based reflex agent. These maintain an internal model of the environment to make decisions considering past states as well as what you’re asking it to do right now.

So you could use a contextual chat bot to develop these chat bots for websites or, or internal PO portals that can maintain conversational context. It can remember previous interactions, and then provide more relevant information or support when the employee or the customer comes back for, for a follow-up or for additional information.

Do sentiment monitoring with that, that historical context. Agents that track media or social media sentiment over time can identify trends and, and give you historical context to current conversations. So you know, something’s being discussed around the organization. It can say, well, you know, two weeks ago this conversation happened then that weighs on what’s going on in these [00:18:00] conversations today.

And then there’s automated information retrieval, uh, agents that can access and synthesize information from internal databases or external sources based on what you ask it. Uh, providing more comprehensive answers than you get from the simple reflex agents. Goal-based agents make decisions to achieve a specific goal, planning a sequence of actions to reach that objective.

This is what most of us think about when we’re thinking of agents, automated press release, distribute distribution, social media, campaign management, internal communication, workflow automation. This is all possible here. I think I, I referenced on an earlier episode that I used an agent, a test agent that I think was Anthropic had set up, and I had it go out to my company’s website, identify our areas of subject matter expertise, and the markets we’re in.

Then go out and find 10. Good podcasts with large audiences where we [00:19:00] could pitch our subject matter experts as guests and it would be an appropriate pitch. And I sat back and watched while it did all of these things. So this is what we’ve got coming. Fourth are utility based agents that choose actions that maximize their utility or a defined performance measure considering various possible outcomes.

Uh, we can use these to optimize communication channel usage, right? Analyze how audiences engage across different communication channels and recommend the most effective platforms for specific messages or, uh, desired reach or desired impact. I can use this for crisis communication, simulation and planning.

Personalized communication delivery. Fifth is learning agents that improve their performance over time by learning from their experiences. You can use this to refine your message targeting, to improve, uh, the, the natural language understanding of chatbots that are engaging with customers or employees or whoever.

And to predict [00:20:00] communication effectiveness. They can analyze a number of factors like message, content, timing, audience demographics. To predict the potential reach and impact of your communications, letting you make adjustments. Sixth are hierarchical agents that break down complex goals into smaller, more manageable sub goals.

Here you’ll have higher level agents overseeing the work of lower level agents, so you’ll have a human manager managing an AI agent who manages AI agents. These for large scale communication projects, multi-channel campaigns, and and streamlining the approval process or use cases. And finally, there are multi-system agents.

These are multiple agents interacting with each other to achieve a common goal or individual goals. Integrated communication, planning and execution. Managing online reputation with agents, monitoring different online platforms, analyzing sentiment, coordinating responses or engagement based on a unified strategy, and then [00:21:00] cross departmental communication coordination.

So we need to understand the distinct capabilities of these different types of agents, and if we do, we’ll be able to leverage them to automate, to gain deeper insights, to do better personalization and better achieve our objectives. And I think, I think this is also a, a, a good point to mention. I have not had a chance to, to read it because you said you saw it and commented on it today.

It’s still early here where I am. But Zora Artis, our interview guest posted something that kind of fits in here too, right?

Neville Hobson (2): Yeah, she shared a post from LinkedIn, which I found quite intriguing. Uh, written by, uh, Jade Beard Stevens, who’s the Director of Digital and Social Innovation at YMU in London. Brief post, but it says it all, I gotta read it out.

It’s quite, quite short. Uh, she says I wasn’t shocked, but still had to share. This rumor has it that open AI is quietly working on a native Shopify checkout. Inside chat. GPT apparently leaked code shows Shopify checkout, [00:22:00] URL Buy Now product offer ratings. No redirects, no search, just chat compare and buy in one flow.

If this happens, Google, TikTok, even product pages as we know them are all about to change. This isn’t just another e-commerce update. This is the merger of search and checkout. This is AI becoming the new storefront. Brands will need to optimize for AI’s first visibility, not just SEO. This could be bigger than TikTok shop, and it’s already happening.

Now, is this a agent ai? I don’t know. Shell, it’s, it’s, it’s kind of fits somewhere in, in this overall picture of, uh, tools, emerging methods emerging. Uh, look at the seven things you, you read out. Uh, there’s some real interesting stuff in there to, to deep dive into, but what Jade mentions is definitely something to pay attention to, even if you’re not in retail or in e-commerce or any of that.

There’s a huge, not huge kind of developing conversation on Reddit about this, which has some more, in more detail on what’s happening. I did a quick search on [00:23:00] this. This is generally this topic to see, you know, anything else talking. I did find something, which isn’t this, this is gonna replace this other thing that I found, I think, which is a Shopify AI chatbot via chat, GPT as the title of the app goes, uh, put out by, um, uh, not, not Shopify beg, pardon?

Shockly. A company called Shockly that, uh, builds, uh, tools to, for, for vendors on Shopify to, to sell their stuff. This isn’t it, but this has been around since September of 2024, and it is actually quite interesting. It’s an app you install. I see it’s got, uh, just under 30, uh, ratings, all five out of five stars from vendors.

Um, it is all to do with, uh, enabling your whole, uh. Storefront using a, a tool from chat chat, GPT. What, um, Jade’s article talks about is this sort of [00:24:00] thing happening natively within Shopify. So that’s a slightly different proposition, but something like this is coming, so you’ve already got third party apps doing this.

Now you’re gonna have a native app doing this. And if it is, um, well, I don’t wanna get hung up on the word digic here, but if, if this is, uh, uh, enables you to, to complete the whole buying process, from interest to purchase, to signing up and paying for it all within chat GPT, that will, uh, a appeal to quite a few people.

I think if it’s offered something better, faster, or less stressful, less hassle, easier than doing it otherwise in, in, uh, in Shopify, it’ll attract attention. So add this one to the list of things to pay attention to as well.

Shel Holtz: Yeah, and whether that’s part of an agent or not, I think depends. It could absolutely be, uh, I could see how that would work in an agent tech environment.

I’m thinking of giving the, the agent the [00:25:00] assignment of buying me a new mirrorless camera, as long as I provide it with the criteria, my price limit of the features that it needs to have, how soon it can be delivered, which brands I don’t want you to consider, uh, but go out and do comparisons of the different models, uh, from different manufacturers that meet my criteria.

Then do price comparison to find the best price. Once you have found the best price, buy it and have it delivered so that I don’t have to do anything else. That’s an agent. So again, you know, if there’s price at the end, what can communicators do with that? I don’t know how much the PR folks can do with that, but the marketing side of the house can probably do a ton with that.

Neville Hobson (2): Yeah. So one more to pay attention to. I was looking through the HubSpot article you referenced, and I, it’s a couple things in there that I, that struck me, uh, their views. Uh, one where they talk about under the, uh, autonomous AI agents paragraph, it’s always a good idea to keep a human involved in any AI operation.

Absolutely [00:26:00] agree with that. Um, a lot of very useful, uh, information in HubSpot’s piece. Uh, some good explainers of what some of this stuff means. And then, um, uh, the answer to the question about preparing for an agent, ai future experimenting. I think the concluding sentence is probably the kind of, okay.

Summarize the whole thing into this. The future is agent. Will you be ready now? That’s what we asked in 4 58 when we talked about this topic, and I wonder if we’ll be asking it again after this one. We’ll see.

Shel Holtz: Undoubtedly we’ll be asking this for some time because even after the agents. Have fully arrived and are available.

Uh, I think there’s going to be a lot of people in our profession and across industry who are not ready

Neville Hobson (2): opportunity for.

Shel Holtz: And we’ll talk about that more when we cover another story later.

Neville Hobson (2): We will. Yeah. So let’s take a look at something quite interesting that popped up in the last few days. [00:27:00] Imagine an AI tool that promises to help you cheat on everything from job interviews to academic exams.

That’s exactly what clearly offers. Created by two former Columbia University students, Chung and Roy Lee and Neil Han Mugham clearly acts as an invisible AI assistant that overlays realtime support onto any application a user is running. It gained attention and controversy after Roy Lee was suspended from Columbia for using an early version during a job interview.

Despite this, clearly has just raised $5.3 million in funding from investors promoting its vision of true AI maximalism, where AI can assist in any life situation without detection. The tool is designed to be undetectable, providing realtime suggestions during interviews, exams, writing assignments, and more, much like an augmented reality layer.

But for conversation and tasks, supporters argue it could level the playing field for those who struggle with traditional [00:28:00] assessments, but critics warn it crosses a serious ethical line, potentially devaluing qualifications and undermining trust in recruitment and academic credentials. Realtime interview assistants raises questions, not just about competence, but about honesty and disclosure.

Rarely happens. Interestingly, the Verge tested it. Their real world testing found that clearly is still very rough around the edges. Technical issues, latency and clunky interactions make it more proof of concept than polished products, at least for now. And did I mention they just got over $5 million in investor funding?

The founders defend the provocative framing. They describe cheating as a metaphor for how powerful AI assistance will soon feel. Much like the early controversies over calculators or spellcheck, as they say, not quite the same thing. I don’t think Shel, but so are we looking at the next Grammarly or are we opening the door to a darker future where nobody can be sure what’s real anymore?

So question for you then Shell is what does this tell us about the [00:29:00] blurring lines between assistance and deception in an AI driven world?

Shel Holtz: Well, I think there’s a couple of ways to look at this. I did hear Lee interviewed on Hard Fork. Uh, it was a great interview and he made a couple of points. First of all, he said that having been through these types of interviews, this is, uh, the kind of interviewing you do for a coding job.

That the tests that they give you have absolutely no relevance to the kind of work that you’re doing. You’re gonna do this once for the interview, and then you’re never gonna do it again. So he doesn’t think that helping people. Figure out how to do that particular exercise is, is all that much of a cheat.

But he also said that everybody programs with the help of AI these days and he says it just doesn’t make sense to have any kind of interview format that assumes you don’t have the use of AI to help you code. I absolutely see that point, but on the other hand, I think this is [00:30:00] just one instance of the kind of thing that AI is going to enable.

And there will be times that it can be very problematic, much more problematic than in this case if somebody can cheat on, say their legal exam or their medical exam, then you’ve got a problem. Somebody who’s not prepared to go out there and and operate on you past the boards because they had help from a program that was written to help them cheat and pass.

So it’s the type of thing that society needs to be thinking about and isn’t yet.

Neville Hobson (2): So if I get this right from what you said, Roy Lee thinks it’s okay to cheat in coding ‘cause it’s a stupid question to ask and you’re only ever gonna do it once. So therefore it’s okay to cheat. Meaning you actually pretend you do know how to do this even though you don’t.

I mean, that is bullshit, frankly, truly. Don’t you think?

Shel Holtz: Well, his his point is that, yeah, you, you don’t know [00:31:00] how to do it, but you don’t have to because you’re never going to on the job.

Neville Hobson (2): So don’t, don’t, don’t, don’t even take the exam and don’t apply for that job. That’s what I would say.

Shel Holtz: I guess then you don’t get any jobs, right?

Well, cheating is

Neville Hobson (2): cheating

Shel Holtz: His point is that you’re, well, yeah, it’s cheating. Yeah. But he says his point is that the cheating in this instance isn’t going to affect your ability to do the job. Whereas in other instances, well, I’m still cheating. I’m not defending it. Understand. I’m just telling you what he said.

Neville Hobson (2): Yeah, sure. Yeah. But it’s still cheating. I, I would say, I mean, it is, to me, this is the same as saying, or someone’s a little bit pregnant or, you know, I’m, I’m, I’m, you know, that kind of stupid kind of defensive argument. This is an indefensible situation in my view that

Shel Holtz: of course, it used to be considered.

Neville Hobson (2): Yeah, but no, no, you can’t. You can’t do it by degrees. She, I don’t believe, honestly, I don’t. You are cheating or you are not. And in this case, again, from how you describe what Roy Lee said, effectively it’s saying, well this is a dumb question to ask and [00:32:00] I’m never gonna do this again, so I’ll get this thing to do it for me basically.

And that they won’t know this. That’s the other thing. They do not know this. They think, are you’s a smart guy? This fell, let’s give him the job. What a ridiculous outcome. And the other ones you mentioned in degrees, you know, taking legal exams or, or you know, passing to be a surgeon. Yeah, they’re serious too, but they’re all the same.

They’re cheating. But I then kind of flip a bit by saying that this is society as we are. I’m afraid this is humans doing this. This will be out there. And this makes it even more difficult to know what’s true and what’s not, and who you can trust and who you can’t. So, you know, welcome to the new world there.

Shel Holtz: I think the adaptation that has to happen has to happen on the part of the people conducting the interviews, not the people taking them. And the reason for that is, I mean, if you think about it, it used to be considered cheating to, to bring a calculator into, well, they mentioned that’s

Neville Hobson (2): the argument he gives.

Ridiculous.

Shel Holtz: Yeah. Well, I mean, everybody’s allowed to use a [00:33:00] calculator now because the people that was 60,

Neville Hobson (2): 60 years ago. Yeah. So maybe in 50 years this would be normal. Yeah.

Shel Holtz: Who conduct the tests came to realize that the people who do the work are able to use calculators. So they should have been part of the test all along.

So I think that’s a legitimate argument, not a, not a legitimate argument for cheating, but for updating the testing so that people don’t feel like they need to.

Neville Hobson (2): So in the meantime, that’s not the landscape. So they need to develop it. So maybe the simplest way to do this is send your AI agent in to take the exam for you.

Has that,

Shel Holtz: well, there are people doing that for job interviews. Yeah, of course. They, they’re probably pretty close to that. Yep. We’ve seen some interesting developments recently with two platforms taking different approaches to verification, and I think some of this may be a little backlash to X, where now you can just buy the blue check mark and it doesn’t actually verify anything other than that you pony up the money for it.

But LinkedIn and Blue Sky [00:34:00] have taken steps with their verification programs. Let’s start with LinkedIn, which is allowing verified identities to extend beyond its own platform. This change means your verified LinkedIn identity can now be visible on other platforms designed to enhance trust and transparency across the internet.

The system leverages open standards and cryptographic methods to ensure authenticity and security. What makes this particularly interesting is how it integrates with Adobe’s technology. Adobe’s content credential system is one of the tools supporting this cross-platform verification. So when you verify your identity on LinkedIn, that verification status can essentially travel with you to other websites and services that support these standards, including Adobe’s Behance.

Now, this is a site that helps creators and people who need to hire creators connect. Now, this is a fundamental shift in how verification works rather [00:35:00] than a siloed verification system on each platform. LinkedIn’s embracing an interoperable approach that lets your verified status function as a digital passport of sorts.

Now, while it’s too bad, this isn’t tied directly to the fedi verse protocols, the significance for communications professionals can’t be overstated. As content creation becomes increasingly distributed across platforms, having a verified identity that travels with you simplifies your ability to establish authenticity in multiple spaces.

For organizations managing multiple spokespersons or content creators, this can streamline verification processes considerably. Meanwhile, blue Sky has taken a different but equally innovative approach to verification by introducing a new Blue Check system just last week. Uh, they’re implementing what they call a user-friendly, easily recognizable blue check mark that will appear next to verified accounts.[00:36:00]

The platform will proactively verify authentic and notable accounts while also allowing trusted verifiers select independent organizations that can verify accounts directly. Now, what’s really interesting about Blue Sky’s approach is how it distributes verification authority. Under this system, organizations like the New York Times can now issue blue checks to their journalists directly within the app, and Blue Sky’s moderation team will review each verification to ensure that it is what they say it is.

This creates a more decentralized verification ecosystem rather than putting all verification power in the hands of the platform itself. Blue Skies verification system has transparency built in. Users can tap on someone’s verified status to see which trusted verifier granted the verification. This adds a layer of context that helps users understand not just that the accounts verified.

But who [00:37:00] vouched for it? Now, before this update, blue Sky had been relying on a domain based verification system letting users set their website as their username. For example, NPR .(JavaScript must be enabled to view this email address) and US Senators verify their account with their senate.gov domains. This method is gonna continue alongside the new blue check mark system, and this gives users multiple ways to establish authenticity.

Now, the evolution of these verification systems comes at a critical time with scammers and impersonators on the rise. A recent analysis found that 44% of the top a hundred most followed accounts on blue sky had at least one doppelganger account attempting to impersonate them. For those of us working in organizational communication, these developments signal a series of important trends.

First. Verification is important and it’s becoming distributed and contextual rather than a single authority declaring who’s authentic. We’re moving toward [00:38:00] ecosystems where multiple trusted entities can vouch for identity. Second Cross platform verification is emerging as a solution to digital fragmentation.

LinkedIn’s approach particularly shows how verified identity could function seamlessly across digital spaces rather than being siloed within individual platforms. Third, transparency about who is doing the verifying is becoming important. Blue Sky’s approach of showing which organization verified an account recognizes that the source of verification matters almost as much as the verification itself.

For organizations, these trends suggest that we really ought to be thinking more holistically about verification strategies. Rather than just get verified on each individual platform, we are really gonna need to start thinking about establishing verified digital identities that can travel with our content and our spokespersons across the net.

Neville Hobson (2): Very interesting development. So I [00:39:00] hadn’t familiarized myself much with the LinkedIn one, but that’s e equally very interesting. Uh, blue Sky though, to me is definitely moving ahead in a very interesting area. Unlike XI think you mentioned Shell, but some people are seeing this as like a slap in the face to Musk.

That’s probably a very tangential way down the, the priority list, but yes, I bet they are. But I found it most interesting the way in which they’ve gone about this in terms of the, the levels of verification. You’ve got your little blue check mark looking slightly differently depending on the verification system.

And by the way, this is, I think it’s a smart move to follow the blue check, although technically it’s not a blue check, it’s a white check in a blue background, but whatever people call it a blue check mark because, uh, it’s familiar thanks to Twitter as was and the who. Trashed it completely. ‘cause the only verification means you’ve paid Musk so many dollars a month fee and therefore you verified.

I mean, that’s Twitter’s def or X’s definition of what verification means. No value to it, in my view, shall frankly. But, uh, this though, [00:40:00] I think is, is far more interesting. Particularly the transparency about who has verified you. Um, I’ve used my own domain, a domain I acquired back in 2023 for the purpose of this is to verify my handle by domain.

Neville Hobson xyz, YXYZ. You might ask because that’s because at the time the Metaverse was a big deal. NFTs were hot, and everyone who was, everyone had a domain ending in X, Y, Z. So hey, that’s a bandwagon I’ll jump onto, which I did. So I’m now using it have been, and it’s only used for that purpose currently.

So, um, you can’t request verification. That’s another thing to mention with Blue Sky, uh, it’s not much you are invited. Is that suddenly that you might get a, not from saying they have verified you or one of these other organizations might, if you excuse me. On a domain with your employer, they can verify you.

And there is something equally interesting on this. I’m not quite sure if this is just a sample, it’ll stay around or not. But you can actually verify yourself. I’ve [00:41:00] seen some people doing that. I haven’t done it. So because I can’t see the point. Uh, ‘cause the point of verification to me is trust in someone else has verified you, not you doing it yourself.

So maybe that will disappear or it’ll have some other function, I don’t know. But the transparency thing, according to the screenshots in Blue Sky’s, uh, announcement posts about this are, are great. A very clear so-and-so is verified. Uh, it says this account has been verified. It has a blue check because it’s been verified by trusted sources.

Then it lists who those sources are and the date they perform. The verification adds lots to the trustworthiness that you perceive rather than just some simply say, yep, you verified, you a blue check. If you’re an organization verify, you’ll have a different style check. And these will all become quite familiar.

They, they’re not complicated at all. So I. You are right at what you said earlier, which is about, verification isn’t just a casual thing anymore. You need to have a strategy about who in your organization, if you are a, a large organization in particular, who gets [00:42:00] verified for what purpose by whom, and we’ll see that emerging as this picks up.

But this is a great start. They do say, and this is going back to the domain, you can self verify with a domain. That’s the only thing that makes sense, because to do it, you’ve got to make changes at your registrar in the in DNS settings and, and a few other things. And also engage with blue sky to do this.

So it’s uh, uh, they say during this initial phase, they’re not accepting direct applications, as I mentioned. Uh, but they do say as this feature stabilizes, so I guess all the excitement’s dying down and people see how it’s all working, they’ll launch a request form for notable and authentic accounts interested becoming verified or becoming trusted verified.

So during the course of 2025, we’ll see this develop and maybe, um, uh, maybe it will, uh, become the kind of benchmark standard for verification on social networks like this. So it’s interesting. I.

Shel Holtz: We need a standard and I’d like to see that [00:43:00] standard. Yeah. Integrated with the fedi verse standards, because these all ought to be infra operable.

We, we really ought to be able to share a post in one place where we are verified and have that post show up wherever people have chosen to follow us from and have that verification show up with us. And people should be able to click on that verification and see who vouched for us. Uh, they should be able to see that the spokesperson for my company was verified by me or by the CEO and it all works together.

Neville Hobson (2): I think that will emerge, uh, thinking about this cross. Posting idea that’s in been in place in a couple of places, but it’s very, very flaky. I’m talking about things like, for instance, it’s been for a while, at least a year, if not longer, where a plugin on WordPress lets you publish your post and it will then share it across, across the Fed us via connection, uh, with Mastodon.

And you’ve then got threads doing the same thing, [00:44:00] but they’re not. They require tweaks to your platform. Uh, the, probably the one that shows you, if I can use this phrase again, the direction of travel is ghost. The, uh, the new platform, which I joined the beginning of this year that has just enabled, um, or recently just enabled the ability to share your posts with Blue Sky Now ghost.

Has invested a lot of time, effort, and probably a bit of money too, I think, into its social web offering, which is in beta. And that’s all to do with the activity pub protocol because Blue Sky has a different protocol at Proto yet that works from ghost to blue sky via a bridge. And that’s a little technical and that has got to be just immediate term usage whilst this, this plays out further.

So someone like Ghost is making big inroads into doing, into enabling this kind of thing. And I would say we’re gonna see a lot of activity [00:45:00] during 2025 from Mastodon in particular, as well as people like Ghost and others to connect up these, these disparate elements of the Fedi verse so that we we’re becoming more cohesive.

But it’s gonna take time.

Shel Holtz: Yeah, the fedi verse is, is nascent, but it’s also, I think, inevitable. We’ve been talking for quite some time now about what is the successor to Twitter now that X has become what it has become. And I’m not sure that there is a successor. I think that there are a number of places that people are attracted to.

It could be ghost, uh, could be for its newsletter functionality as much as for its blogging functionality. It could be threads it, it could be blue sky, it could, you know, whatever. But as long as where I am, I can follow who I want to follow and have that appear in the network that I have chosen, I’m good.

So I think this is where things are, are headed inevitably since I think the days of somebody being able [00:46:00] to come along and say, I’m the new 800 pound gorilla of social networking. Everybody’s coming here are over. I.

Neville Hobson (2): Yeah, it’s been apparent like that, that that’s likely to be the case for, for a bit, I believe very much that the time is gone for, for monolithic centralized social networks like Facebook, for instance.

No, this is the time for niche networks. Uh, people can set things up themselves. Uh, it doesn’t matter. You, you’ve got 50 people on there or 50,000 people on there, doesn’t matter. And indeed, the the recent, uh, outage on Blue Sky is a, is an interesting indicator of the fragility of all of this. And, and Dan’s gonna talk about this a bit later in his report, but this is an interesting time.

We’re now, it’s almost like things are maturing, it seems to me. And I think you’re right when you say that, that people aren’t, aren’t so much attracted by the idea of a centralized place where, Hey, we’ve all gotta go here after the experience on X. You’ve got more about people saying, I want to get outta here, where do I go?

So, um, we’re still at that phase, and you’ve got. Something interesting with Trump’s, uh, not Trump Musk’s, uh, GR [00:47:00] network, developing chatbots for it and all this stuff. So that’s something interesting in that area of this. So it’s all at a time for communicators to pay attention closer to what is happening here and the implications of it just as you and I are doing.

And if you don’t wanna do that, that’s fine. Just listen to FIR ‘cause we’ll help you understand it.

Yep. Okay. That’s a really good report, Dan. Thank you. Good topics. You’ve talked about, uh, blue sky. I mentioned just before your report actually the outage was unfortunate, but is it not an indicator precisely of that fragility? I mentioned previously different definitions of decentralization that you mentioned.

I think that’s. Possibly a communication issue because people seem to be latching onto, Hey, it’s decentralized when actually it’s more like, it’s going to be decentralized. ‘cause that’s our aspiration that we’re working towards, which is the case with, with Blue Sky. That’s very good on threads move to.com and web improvements.

I must admit, I, I was a bit yawny about [00:48:00] that. You know, dot net.com. Do I care as a user? Well maybe I should because I then read somewhere else that the move to.com enables meta to do things that they can’t do with ANet domain. And I’m sure you’ll know more about that than me Dam at the internet Society.

Again, interesting developments with what’s happening with all of this. So thanks for the report, Dan. This is really, really a good one.

04/28/25 | 1 Comment | FIR #462: Cheaters Never Prosper (Unless They’re Paid $5 Million for Their Tool)

Comments

1.What a crazy story. I love the usefulness of AI, but of course alongside with productivity, people will use it for taking advantage.

Cheers, Shel

Oregon Tim | May 2025 | Portland

« Back

Contact

Email: .(JavaScript must be enabled to view this email address)
Phone: +1.415.881.7435
Cell: 925.864.2249
Skype: shelholtz
Twitter: @shelholtz

E-Newsletter Sign Up

Nuzzel has discontinued its newsletter service (along with everything else). I am actively seeking a new solution to continue producing a short, useful newsletter. Stay tuned. In the meantime, you are welcome to visit my link blog, from which I choose the stories I include in my newsletter.